Real analysis — rigorous epsilon-delta, metric spaces

4.10.22 · Maths › Advanced Topics (Elite Level)

1. The epsilon-delta definition of a limit

WHAT each piece means

$∣ f (x) - L ∣ < ε$ : output target tube mein hai, jiska half-width $ε$ hai $L$ ke around.
$0 < ∣ x - a ∣ < δ$ : input $a$ ki ek punctured neighborhood mein hai (hum $x = a$ ko ignore karte hain — limit neighbourhood ke baare mein hai, point ke baare mein nahi).
Quantifiers ka order sacred hai: $ε$ pehle aata hai (challenge), phir $δ$ depend kar sakta hai $ε$ pe.

2. Deriving a $δ$ from scratch (the working method)

$lim_{x \to a} f (x) = L$ prove karne ke liye tumhe $δ$ ko $ε$ ke function ke roop mein produce karna hoga. Strategy:

Goal $∣ f (x) - L ∣ < ε$ se shuru karo.
$∣ x - a ∣$ ko factor out karo: likho $∣ f (x) - L ∣ = ∣ x - a ∣ \cdot (stuff)$ .
Ek chhoti neighborhood pe "stuff" ko ek constant $M$ se bound karo (iske liye aksar ek preliminary restriction chahiye hoti hai jaise $∣ x - a ∣ < 1$ ).
Tab $∣ f (x) - L ∣ < M ∣ x - a ∣ < ε$ hold karta hai jab $∣ x - a ∣ < ε / M$ .
Choose karo $δ = min (1, ε / M)$ . Done.

Figure — Real analysis — rigorous epsilon-delta, metric spaces

3. Continuity (limit that hits the point)

4. Metric spaces — abstracting "distance"

Standard metrics

Space	Metric $d (x, y)$	Name
$R$	$∣ x - y ∣$	usual
$R^{n}$	$\sum (x_{i} - y_{i})^{2}$	Euclidean
$R^{n}$	$\sum ∣ x_{i} - y_{i} ∣$	taxicab/ $ℓ^{1}$
$R^{n}$	$max_{i} ∣ x_{i} - y_{i} ∣$	Chebyshev/ $ℓ^{\infty}$
any set	$1$ if $x \neq = y$ , else $0$	discrete

5. Forecast-then-Verify

Recall Aage padhne se pehle forecast karo

Q: $lim_{x \to 4} x = 2$ ke liye, ek kaam karne wala $δ (ε)$ predict karo. Verify: $∣ x - 2∣ = \frac{∣ x - 4∣}{x + 2} \leq \frac{∣ x - 4∣}{2}$ (kyunki $x \geq 0$ ). Toh $∣ x - 4∣ < 2 ε$ kaam karta hai: $δ = 2 ε$ . Divide kyun? $x - 2$ ko rationalise karo $\frac{x + 2}{x + 2}$ se multiply karke taaki $∣ x - 4∣$ expose ho.

Flashcards

lim_{x \to a} f (x) = L

ki rigorous definition

\forall ε > 0 \exists δ > 0 : 0 < ∣ x - a ∣ < δ \Rightarrow ∣ f (x) - L ∣ < ε

Kaun sa quantifier pehle aata hai,

ε

δ

, aur kyun

ε

pehle;

δ

depend kar sakta hai

ε

pe (challenge phir response)

Limit mein "

0 < ∣ x - a ∣

" (punctured) kyun hota hai

Limit

x = a

ko ignore karti hai; yeh

a

ke paas ke behaviour ke baare mein hai,

a

par value ke baare mein nahi

δ

find karne ki general recipe

∣ f (x) - L ∣ = ∣ x - a ∣ \cdot M

factor karo,

∣ x - a ∣ < 1

se cap karo, phir

δ = min (1, ε / M)

lim_{x \to 3} x^{2} = 9

ke liye

δ

δ = min (1, ε /7)

Difference: continuity vs uniform continuity

Continuity mein

δ

point pe depend kar sakta hai; uniform continuity ko sab points ke liye ek

δ

chahiye

Teen metric axioms

d (x, y) = 0 ⟺ x = y

; symmetry; triangle inequality

Discrete metric

d (x, y) = 0

agar

x = y

, warna

1

Open ball

B (a, r)

ki definition

{x : d (x, a) < r}

Metric space mein limit

\forall ε \exists N : n \geq N \Rightarrow d (x_{n}, L) < ε

Euclidean triangle inequality prove karne wali inequality

Cauchy–Schwarz,

x \cdot y \leq ∥ x ∥∥ y ∥

Recall Feynman: ek 12-saal ke bacche ko explain karo

Ek dartboard imagine karo. Ek dost kehta hai "main bet lagata hoon ki tum bullseye ke 1 cm ke andar nahi maar sakte." Tum kehte ho "theek hai, bas mujhe board ke paas khadha hone do." Agar har distance ke liye jo woh pick kare — 1 cm, 1 mm, yahan tak ki ek baal ki chaurai — tum hamesha ek aisi jagah dhundh sako jahan khade hone se guarantee ho ki itne paas maaro, toh hum kehte hain tumhare darts sach mein bullseye approach karte hain. Yahi limit hai. $ε$ hai kitna close woh demand karte hain; $δ$ hai tum kitna paas khade hote ho. "Metric" bas koi bhi fair ruler hai do cheezein kitni door hain yeh measure karne ke liye, jab tak ruler honest ho: zero distance sirf identical cheezein ke liye, dono taraf same, aur koi shortcut seedha jaane se behtar nahi (triangle inequality).

Connections

Limits and Continuity — woh calculus jise yeh rigorous banata hai
Sequences and Series — convergence $N$ – $ε$ version hai
Cauchy Sequences and Completeness — jab limits space ke andar exist karti hain
Topology — open sets kisi bhi metric se generalize ho jaate hain
Cauchy–Schwarz Inequality — Euclidean triangle inequality ka engine
Uniform Continuity and Compactness — Heine–Cantor theorem